On Thu, 12 Sep 2024 11:14:32 GMT, Per Minborg wrote:
>>> Wdyt about the benchmark I have added at [#20829
>>> (comment)](https://github.com/openjdk/jdk/pull/20829#issuecomment-2326404582)?
>>> Sadly I didn't yet run against this PR
>>
>> Let's see what we can do about this later.
>
>> Wdyt abo
On Fri, 6 Sep 2024 12:05:21 GMT, Per Minborg wrote:
> Wdyt about the benchmark I have added at [#20829
> (comment)](https://github.com/openjdk/jdk/pull/20829#issuecomment-2326404582)?
Added https://bugs.openjdk.org/browse/JDK-833
-
PR Comment: https://git.openjdk.org/jdk/pull/
On Tue, 3 Sep 2024 07:52:44 GMT, Per Minborg wrote:
> This PR proposes to handle smaller FFM copy operations with Java code rather
> than transitioning to native code. This will improve performance. In this PR,
> copy operations involving zero to 63 bytes will be handled by Java code.
>
> Here
On Tue, 3 Sep 2024 07:52:44 GMT, Per Minborg wrote:
> This PR proposes to handle smaller FFM copy operations with Java code rather
> than transitioning to native code. This will improve performance. In this PR,
> copy operations involving zero to 63 bytes will be handled by Java code.
>
> Here
On Tue, 3 Sep 2024 15:23:20 GMT, Maurizio Cimadamore
wrote:
>> src/java.base/share/classes/jdk/internal/foreign/AbstractMemorySegmentImpl.java
>> line 637:
>>
>>> 635: for (; offset < limit; offset += 8) {
>>> 636: final long v =
>>> SCOPED_MEMORY_ACCESS.getLong(sr
On Tue, 3 Sep 2024 15:44:34 GMT, Maurizio Cimadamore
wrote:
>> src/java.base/share/classes/jdk/internal/foreign/AbstractMemorySegmentImpl.java
>> line 642:
>>
>>> 640: // 0...0X00
>>> 641: if (remaining >= 4) {
>>> 642: final int v =
>>> SCOPED_MEMORY_A
On Tue, 3 Sep 2024 11:51:57 GMT, Francesco Nigro wrote:
>> This PR proposes to handle smaller FFM copy operations with Java code rather
>> than transitioning to native code. This will improve performance. In this
>> PR, copy operations involving zero to 63 bytes will be handled by Java code.
>>
On Tue, 3 Sep 2024 15:22:36 GMT, Maurizio Cimadamore
wrote:
>> This PR proposes to handle smaller FFM copy operations with Java code rather
>> than transitioning to native code. This will improve performance. In this
>> PR, copy operations involving zero to 63 bytes will be handled by Java cod
On Tue, 3 Sep 2024 07:52:44 GMT, Per Minborg wrote:
> This PR proposes to handle smaller FFM copy operations with Java code rather
> than transitioning to native code. This will improve performance. In this PR,
> copy operations involving zero to 63 bytes will be handled by Java code.
>
> Here
On Tue, 3 Sep 2024 11:32:26 GMT, Maurizio Cimadamore
wrote:
>> This PR proposes to handle smaller FFM copy operations with Java code rather
>> than transitioning to native code. This will improve performance. In this
>> PR, copy operations involving zero to 63 bytes will be handled by Java cod
On Tue, 3 Sep 2024 07:52:44 GMT, Per Minborg wrote:
> This PR proposes to handle smaller FFM copy operations with Java code rather
> than transitioning to native code. This will improve performance. In this PR,
> copy operations involving zero to 63 bytes will be handled by Java code.
>
> Here
On Tue, 3 Sep 2024 07:52:44 GMT, Per Minborg wrote:
> This PR proposes to handle smaller FFM copy operations with Java code rather
> than transitioning to native code. This will improve performance. In this PR,
> copy operations involving zero to 63 bytes will be handled by Java code.
>
> Here
On Tue, 3 Sep 2024 07:52:44 GMT, Per Minborg wrote:
> This PR proposes to handle smaller FFM copy operations with Java code rather
> than transitioning to native code. This will improve performance. In this PR,
> copy operations involving zero to 63 bytes will be handled by Java code.
>
> Here
On Tue, 3 Sep 2024 07:52:44 GMT, Per Minborg wrote:
> This PR proposes to handle smaller FFM copy operations with Java code rather
> than transitioning to native code. This will improve performance. In this PR,
> copy operations involving zero to 63 bytes will be handled by Java code.
>
> Here
This PR proposes to handle smaller FFM copy operations with Java code rather
than transitioning to native code. This will improve performance. In this PR,
copy operations involving zero to 63 bytes will be handled by Java code.
Here is what it looks like for Windows x64:
